Professional Documents
Culture Documents
public domain and may be used and reprint- To obtain copies of this report and
ed without special permission; citation as to
other BTS products, go to the online
source is required.
Product Catalog at:
This document is disseminated under the www.bts.gov
sponsorship of the Bureau of Transportation
Statistics, U.S. Department of Transportation You may also mail orders to:
in the interest of information exchange. The Product Orders
U.S. government assumes no liability for its Bureau of Transportation Statistics
contents or use thereof. U.S. Department of Transportation
400 7th Street SW, Room 3430
Washington, DC 20590
Recommended citation phone 202-366-DATA
U.S. Department of Transportation fax 202-366-3640
Bureau of Transportation Statistics email orders@bts.gov
Bicycle and Pedestrian Data:
Sources, Needs, & Gaps To obtain transportation-related
BTS00-02 statistical information, call
Washington, DC: 2000 800-853-1351.
ii
Acknowledgments
Susan J. Lapham
BTS Publications Staff
Associate Director for Marsha Fenn
Statistical Programs and Services Chip Moore
Martha Courtney
1 INTRODUCTION ........................................................................................7
Study Approach..........................................................................................................8
Why Collect Bicycle and Pedestrian Data? ................................................................8
iv
Secondary Data ........................................................................................................59
References ................................................................................................................60
Box
4-1 Portland, Oregon, Pedestrian Potential and Deficiency Indices ......................56
APPENDICES
v
Executive Summary
1
2 Bicycle and Pedestrian Data
This review of existing data and data needs was based on outreach to user
groups that included planners, advocates, and researchers at federal, state,
and local government agencies, universities, and nonprofit organizations.
Published materials and information from other recent assessments of trans-
portation data needs were also reviewed. Primary data sources were classi-
fied by four types: 1) usage, trip, and user characteristics; 2) user preferences;
3) facilities; and 4) crash and safety data. Key types of secondary data (data
that are based on analysis of primary data) were also identified, including
research-study results and manuals of practice.
Based on this review, priorities for data needs were identified based on the
following criteria:
The study also identified a number of opportunities for improving the qual-
ity of bicycle and pedestrian data. Emerging information technology, includ-
ing intelligent transportation systems (ITS) for data collection as well as
information management tools such as geographic information systems
(GIS), have the potential to make data collection, management, and analysis
cheaper and easier. Greater coordination of efforts, however, is required to
fully exploit the potential of these technologies to improve pedestrian and
bicycle data. Also, certain types of data, such as numbers of trips by facility
and user type, are potentially useful to a wide range of user groups; but coor-
dination among these groups is required to establish standardized, mutually
beneficial data collection procedures.
Some of the key recommendations emerging from this study are identified
below. These are areas in which federal involvement—through coordination
of efforts, provision of technical assistance, or sponsorship of research and
development activities—could greatly assist in benefiting a broad range of
user groups.
Executive Summary 3
Table 1
Assessment of Data Priorities
Quality of Priority for
Type of data and description existing data better data
User preferences
Relative preferences for facility
design characteristics and other
supporting factors Fair Medium
Facilities data
Characteristics relating to quality
for bicycle or pedestrian travel Fair Medium
Secondary data
Safety and demand impacts of
design features Fair High
Safety and demand impacts of
policies, programs Fair Medium
General/Cross-Cutting Recommendations:
● Bring together the full range of users of bicycle and pedestrian data to dis-
cuss how specific data collection efforts could benefit the broadest num-
ber of users.
Recommendations on Facilities:
● Facilitate discussions among various data user groups to identify key net-
work characteristics relevant to bicycle and pedestrian planning and
design, and provide guidance to state and local agencies responsible for
collecting and maintaining transportation data.
Executive Summary 5
1
T he importance of pedestrian and bicycle travel has received
increasing attention in recent years as planners and policymakers
recognize the benefits to communities, public health, economic devel-
opment, and the environment of improving nonmotorized travel
options. The Transportation Equity Act for the 21st Century (TEA-
21), passed by Congress in 1998, requires the consideration of bicy-
cle and pedestrian needs in transportation planning and increases
funding opportunities for bicycle and pedestrian projects. TEA-21
also directs the Bureau of Transportation Statistics (BTS) to develop
data on bicycles and pedestrians.
Efforts to plan for bicycle and pedestrian travel are frequently ham-
pered, however, by deficiencies in data on travel characteristics, facilities,
safety, and user preferences. As a step toward improving the quality of
bicycle and pedestrian data, BTS has undertaken this assessment of bicy-
cle and pedestrian data needs. The study has the following objectives:
● First, to provide an inventory of existing sources of bicycle and
pedestrian-related data, including the extent, quality, and limita-
tions of these sources;
● Second, to identify and prioritize areas in which additional or
improved data are needed; and
● Third, to identify and recommend opportunities for improving the
quality of bicycle and pedestrian data.
7
8 Bicycle and Pedestrian Data
STUDY APPROACH
The following methods were used to identify existing sources of bicycle and
pedestrian data, data needs and priorities, and opportunities for improving data
collection:
● Interviews and discussions with key people involved with pedestrian
and/or bicycle issues at a national level;
● An email questionnaire sent to numerous individuals and groups, includ-
ing national, state, and local pedestrian and bicycle planners, advocates,
and researchers;
● Various written sources of information, both published and unpublished;
● Experience gained from previous pedestrian and bicycle projects under-
taken by the authors of the report.
Appendix A provides an overview of the email survey methodology and results.
Appendix B contains a brief review of other recent assessments of data needs.
Research
Research studies may rely wholly or in part on data collected specifically for the
study being conducted. National or local data sources may also be utilized,
Introduction 9
depending on the scope of the study and the quality of available sources. Bicycle
and pedestrian research studies fall into at least three major categories:
● Studies on basic operational characteristics (e.g., bicyclist riding speeds,
shared-path “level-of-service” characteristics, or pedestrian gap accept-
ance at intersections). These studies usually rely on specific data-collec-
tion efforts.
● Studies of factors influencing crashes and safety. These studies rely on at
least two types of data: 1) crash characteristics, describing the people and
vehicles involved, location, environmental conditions, contributing fac-
tors, injury outcomes, etc.; and 2) usage data, including flows by location,
user characteristics, etc., from which measures of exposure and therefore
risk can be derived.
● User preference and demand studies, which analyze the effects of facility
design and other policies on user preferences and demand (usage). These
studies frequently rely on surveys of existing and potential pedestrians
and bicyclists. Quantitative models of behavior can also be used to devel-
op information on user preferences. These models may be based either on
“stated preference” survey data, in which people are asked to make choic-
es among various alternatives, or on “revealed preference” data, i.e.,
observations of people’s actual behavior from travel surveys and counts.
These studies may also utilize trend data on usage, which is analyzed to
determine the impacts of facility or policy changes.
● As researchers synthesize and analyze raw data from research studies, the
results yield secondary data, such as evidence on the impacts of facility
design and other policies on safety, user satisfaction, demand, medical
costs, and other factors. The results may be in the form of direct state-
ments of impact, or they may take the form of coefficients, factors, etc.
that can be used in modeling the joint impacts of multiple actions.
Research results are also used to develop manuals of practice. These secondary
forms of data include design manuals, policies, and programs that, based on the
results of a variety of research studies as well as other factors, can be defined as
“recommended practices.”
Planning and design for pedestrian and bicycle travel involves at least the fol-
lowing aspects:
● The siting and design of facilities for travel (e.g., roads, signals, paths, and
lanes);
10 Bicycle and Pedestrian Data
Policymaking
Information on existing conditions and trends in usage, crash rates, and facili-
ties can provide important background for setting policy and for making fund-
ing and programmatic decisions. These data can help identify whether existing
policies and programs are successful and whether additional or revised policies
and programs are needed. For example, data indicating that walkway quality is
generally poor and declining might lead to the establishment of improvement
programs or funds, or the establishment of a committee to provide input into
improvements. Data showing increasing crash rates may indicate the need for
additional funding for efforts to improve roadway safety.
Data on forecast usage can also be important for policymaking. Demand fore-
casts can be used to predict potential increases in pedestrian and bicycle travel
resulting from various types of improvements. These forecasts can help indicate
the potential benefits and impacts of various policy and programmatic deci-
sions.
The remainder of this document assesses the state of existing data, priorities for
data needs, and opportunities for improving data collection.
Existing Data Sources
CHAPTER
2
T he following sections provide an overview of existing U.S. sources
of pedestrian- and bicycle-related data, including characteristics,
uses, limitations, and potential enhancements of each source. While
the focus is mainly on data sources available at the national level, the
general existence and quality of data sources at the state and local lev-
els are also discussed. For purposes of this report, both primary and
secondary data sources are identified. Primary data sources are clas-
sified by the following types of data:
● Usage, trip, and user characteristics, including couts of users, trip
patterns, and demographic socioeconomic characteristics;1
● Data on the preferences, needs, and attitudes of current or poten-
tial bicyclists and pedestrians;
● Facility characteristics, including the locations and attributes of
pedestrian and bicycle facilities;
1 A separate category could be established on traffic flow characteristics that would include
pedestrian and bicycle volumes, speeds, and traffic densities on individual facilities. Traffic
flow data are routinely collected for motor vehicles to characterize system performance and
congestion, but with the exception of specific research studies, only volume (count) data are
collected for bicycle and pedestrian traffic. In this report, counts are discussed in conjunction
with related survey data on usage, since both are collected with similar objectives.
13
14 Bicycle and Pedestrian Data
● Crash and safety data, including crashes and other data related to the
safety and security of bicycle and pedestrian travel; and
● Expenditures on and capital stocks of vehicles and facilities.
While only primary data sources are treated in detail, a number of secondary
data sources are also discussed briefly because these were identified in the
course of the study as being important to practitioners. These secondary sources
include:
● Results of research studies,
● Manuals of practice, and
● Other sources.
A summary of existing data sources is provided in table 2-1.
This broad category of data answers the questions of who is traveling, how,
where, when, and why. Usage data may take the form of number of travelers
by facility, time of day, geographic area, etc. Data on trips include items such as
origin and destination, trip length, mode, route, purpose, and time of day.
Characteristics of the trip maker or “system user” may include demographic
and socioeconomic factors such as age, gender, income, household structure,
and other information that describes the person and his or her situation in life.
Data may be reported either in aggregate or disaggregate form. Aggregated data
show statistics such as mode split or average income for all persons living in a
census tract. Disaggregated data include records of the characteristics and trip
patterns of individual people or households. Disaggregated data can be ana-
lyzed directly for modeling purposes. Data on usage, trip, and user characteris-
tics are initially collected at a disaggregated level but are frequently reported or
available for analysis only at some level of aggregation. In the aggregate, per-
sonal and trip variables may be cross-classified, e.g., trip length distributions by
mode, purpose, and/or age.
Data sources for usage, trip, and user characteristics include:
● Counts of bicyclists or pedestrians,
● The U.S. decennial census,
● Metropolitan household travel surveys,
● The Nationwide Personal Travel Survey (NPTS),
● Other surveys conducted sporadically at a national level, and
● Various local surveys and market studies.
Table 2-1
Existing Sources of Bicycle and Pedestrian Data (includes national and multistate-level sources only)
15
Continued on next page
16
Bicycle and Pedestrian Data
Table 2-1 (Continued)
17
Continued on next page
18
Table 2-1 (Continued)
GENERAL SOURCES
Data source Agency Scale Frequency Coverage Contents Uses Availability
National U.S. DOT, Bureau National Annual Not Various summary Conditions and trends Published
Transportation of Transportation applicable statistics analysis www.bts.gov
Statistics Statistics
19
20 Bicycle and Pedestrian Data
Counts
Counts indicate how many people or vehicles are using the system at any given
location. Counts may be differentiated by user characteristics, time of day, or
day of week. In addition, if an appropriate sample is conducted, counts can be
summarized to provide estimates of overall facility or system usage by time,
geographic area, facility type, or user type. Counts have a variety of uses,
including:
● Tracking trends in usage,
● Developing exposure measures for crash analysis,
● Evaluating level of service on a facility by comparing volumes to capacity,
● Identifying and prioritizing improvements, and
● Calibrating travel demand models.
Motor vehicle traffic volumes are monitored through the Highway Performance
Monitoring System (HPMS). The HPMS is a national system for tracking
motor vehicle traffic volumes by vehicle type, facility type, and geographic area
based on counts at a carefully selected sample of locations. No equivalent of the
HPMS exists for monitoring pedestrian and bicycle volumes or distance trav-
eled. While a few cities and metropolitan planning organizations routinely con-
duct pedestrian and bicycle counts, most collect them only sporadically for
specific studies or do not collect them at all.
Existing local count data also vary in quality. When designing a study, consid-
eration must be given to choosing sampling times and collecting enough sam-
ples to account for daily, weekly, and seasonal variations, as well as for random
variations resulting from weather, special events, and other causes. If counts are
to be used to estimate traffic in a larger geographic area, a representative set of
locations must also be sampled. A national compendium of local counts was
published in 1994 by the Federal Highway Administration (FHWA) as a sup-
plement to the National Bicycling and Walking Study, although such a com-
pendium is not updated on a regular basis (USDOT FHWA 1994).
Census
Because of the large amount of data that must be collected in the census and the
need to obtain a high level of response, it is unlikely that the census can be mod-
ified to collect more comprehensive data related to pedestrian and bicycle trav-
el. It is possible, however, that further detailed analyses of the existing
disaggregated census data—performed confidentially and released in aggregate
or summary form for public use—could help relate journey-to-work character-
istics to personal characteristics, household characteristics, and any relevant
tract-level data such as population density.
● Walk trips (particularly short trips such as from work to a cash machine)
are commonly underreported. Also, walk trips to access other modes,
notably transit, are often not tracked separately in travel surveys.
● The surveys provide only limited information on the travel patterns of
children. The parent completing the survey is generally asked to report
trips for children from age five to somewhere in the teens. As a result of
this proxy reporting, not all trips may be noted.
MPOs are developing more sophisticated travel survey methods as they search
both for better data and cheaper data-collection methods, but significant limi-
tations still exist. Some potential enhancements to address the above limitations
for pedestrian and bicycle analysis include:
● The survey sample size can be increased to obtain data on more bicycle
and walk trips for analysis. The costs of obtaining a sufficient sample size
for detailed analysis, however, may be prohibitive. As an alternative, areas
with high levels of pedestrian and bicycle travel may be oversampled, or
these groups may be targeted in other ways similar to obtaining sufficient
samples of transit riders through on-board surveys. Portland Metro, for
example, collected additional surveys in densely populated urban neigh-
borhoods that were likely to have higher numbers of pedestrian trips.
Effective means of obtaining larger, representative samples of bicycle and
pedestrian travelers is an area in which further research and implementa-
tion experience is needed.
● Proper survey design and interviewer training can greatly reduce the
underreporting of short-walk trips and transit-access trips.
● A few areas are testing “time-use” surveys, in which all daily activities
rather than just trips are recorded. Time-use surveys, if designed proper-
ly, could identify the time and length (although not the route taken) for
purely recreational walk and bicycle trips. Portland, OR, has adopted this
methodology and it is being tested in Seattle, WA.
● Underreporting of short trips, walk trips, and trips by children may be
reduced or eliminated through the use of global positioning system (GPS)
receivers to record travel patterns. GPS could also be used to identify spe-
cific routes taken for walk and bicycle trips. The use of GPS in travel sur-
veys is currently being explored in Atlanta, GA, Phoenix, AZ, and
Lexington, KY. The first two are multimodal surveys, including bicycle
and pedestrian modes.
● In addition to collecting trip data, household surveys may include a
stated-preference component that can be used to assess travelers’ prefer-
ences for various hypothetical mode and route choices. Stated-preference
components have been added to surveys in areas such as Portland, OR,
24 Bicycle and Pedestrian Data
and Dallas, TX, to assess factors such as road pricing and residential loca-
tion choice. Questions could also be developed that focus on bicycle and
pedestrian travel incentives and options. Conducting stated preference
surveys in conjunction with travel surveys provides compatible data sets
and thus, potentially, more useful information than conducting these sur-
veys independently. On the other hand, creating longer surveys increases
the cost of the survey and also has the potential to reduce the quality and
rate of responses.
Box 2-1
U.S. Consumer Product Safety Commission Study
In 1990 and 1991, the U.S. Consumer Product use roughly double those of most other age
Safety Commission (CPSC) undertook a major groups.
national study of bicycle use and hazard pat-
terns. CPSC conducted two surveys: an injury Questions about riding habits showed that
survey utilizing National Electronic Injury most bicyclists (64 percent) ride a substantial
Surveillance System (NEISS) data and a nation- proportion of the time on neighborhood streets
al random-digit-dial telephone survey to obtain with low traffic volumes. This compares with
exposure data. The results were combined to 29 percent on sidewalks and playgrounds, 17
evaluate bicycle usage and risk factors. percent on bike paths, 18 percent on unpaved
roads, 7 percent on major thoroughfares, and
The exposure survey estimated that there are 11 percent on unpaved surfaces or trails. The
about 67 million bicyclists who ride a total of survey also noted that collisions or near-colli-
about 15 billion hours annually. Most bicycling sions with moving motor vehicles accounted
is for recreational purposes, but almost 9 per- for only about 10 percent of injuries. The
cent of riders use their bicycles primarily for remainder resulted from a variety of other fac-
commuting to work or school. The survey con- tors, such as collisions with other bicyclists,
firms the importance of children’s travel and pedestrians, or stationary objects; stunts; and
recreation, which is not measured in many sur- falls on rough or slippery riding surfaces.
veys: about 22 percent of bicyclists are under
age 10 and 40 percent are under age 15. The SOURCE: Gregory B. Rodgers et al., Bicycle Use and
survey also noted a higher risk of injury among Hazard Patterns in the United States, prepared for the U.S.
Consumer Product Safety Commission, 1994.
children, with injury rates per million hours of
perceived safety. The survey was distributed over the Internet and via
mailed responses to advertisements and, therefore, does not represent a
random sample of commuters.
● Recreational surveys. The National Sporting Goods Association conducts
an annual national survey to determine participation (six or more times
in the previous year) in a variety of sporting activities, including bicycling,
by age and gender. Results are published in the Census Bureau’s annual
Statistical Abstract of the United States. Rodale Press occasionally con-
ducts surveys of 1,000 or more adults to identify participation in bicy-
cling, walking, or running activities according to personal characteristics,
incentives, and availability of facilities. These surveys may be most useful
for tracking national trends in recreational bicycling and walking. Survey
results are published in aggregate format, but this study did not find any
instances where disaggregated data were made publicly available for
analysis.
● Health and physical activity surveys. The Behavioral Risk Factor
Surveillance System (BRFSS) is an annual survey designed to assess
health-related risk factors including physical activity. The BRFSS is
designed by the Centers for Disease Control and Prevention (CDC) and is
administered at the state level by State Departments of Health via month-
ly random samples. The survey is designed with a number of optional
modules that states can implement, including a module on exercise that
includes the frequency and distance of jogging and walking, bicycle hel-
met use, and other information. Survey results could be used to analyze
trends in recreational activity, including analysis by demographic charac-
teristics. Annual sample sizes usually fall within the range of 1,000 to
4,000 per state, allowing some amount of disaggregation within each
state. The modular design of the survey also means that uniform data can
be collected in all states that choose to apply the optional modules. To
date, however, it does not appear that any states have applied the exercise
module. BRFSS survey data can be obtained on CD-ROM from the
Government Printing Office and summary statistics are also available on
the Internet.
Another annual health-related survey is the National Health Interview
Survey (NHIS), which includes the frequency of physical activity as one of
its data items. This survey could also be used for trend analysis. The survey
does not currently ask for type of physical activity, however, and does not
yield data at the level of the local community or metropolitan area.
● NHTSA Survey on Public Beliefs. During 1999, the National Highway
Traffic Safety Administration (NHTSA) conducted a national telephone
survey of 4,000 people on beliefs and awareness about pedestrian and
Existing Data Sources 27
Local Surveys
In some areas local surveys have been conducted of bicyclists and/or pedestri-
ans to assess personal and trip characteristics. University researchers, city or
metropolitan planners, or other interested groups may conduct such surveys.
The content, scope, and quality of these surveys varies considerably according
to the specific purpose, budget, and level of knowledge of those responsible for
the survey. Nevertheless, these surveys can yield useful information for local
planning. Surveys of bicyclists to determine usage patterns and personal prefer-
ences have been conducted by city or regional transportation planning agencies
in areas such as Boulder, CO, Madison, WI, San Diego, CA, and Seattle, WA.
In many cities with rail transit systems, transit agencies routinely conduct pas-
senger surveys that include mode of access to the station. For example, surveys
conducted by the Massachusetts Bay Transportation Authority (MBTA) in
Boston; the Chicago Transit Authority and METRA commuter rail in Chicago;
and the Bay Area Rapid Transit District (BART) in San Francisco have identi-
fied the numbers and percentages of transit accessing each station by walking,
bicycling, or other modes of travel.
Local survey results might in some cases be useful to planners in other areas.
Carefully performed surveys can yield valid and transferable information on
traveler characteristics, trip patterns, and preferences. Some local surveys have
been publicized nationally, for example through FHWA’s National Bicycling
and Walking Study, which published a synthesis of survey findings in 1991
(USDOT FHWA 1992). There is no ongoing effort to identify and disseminate
local survey results, however, and such an undertaking would probably prove
difficult. Furthermore, even if local results could be made available on a wide-
spread basis, it is important to recognize the specific situation and limitations
of each survey before generalizing to other areas.
Attitudinal Surveys
Attitudinal and preference data are often collected in conjunction with surveys
to determine other data, such as usage patterns, personal characteristics, and
crash experience. National survey efforts that investigate preferences, needs,
and attitudes are described above and include the recent national adult bicyclist
survey (Moritz 1997), the 1999 NHTSA Survey on Public Beliefs, and opinion
polls such as the 1991 Harris Poll. Also, as identified above, local surveys have
been conducted in some areas to assess user preferences, needs, and attitudes.
These surveys can be useful for identifying specific concerns and needs in a com-
munity or region.
Modeling
Choice or demand models have been developed by some researchers and plan-
ners to predict bicycle or pedestrian mode or route choice based on character-
istics of the mode or route (e.g., travel time, facilities available). These efforts
have produced some quantitative information on relative preferences, for
example, for bicycle lanes versus off-street trails. The Regional Transportation
Authority in Chicago, for example, recently developed a model to predict mode
of access to transit as a function of sidewalks, pathways, intersection improve-
ments, and bicycle parking facilities in station areas (Wilbur Smith Associates
1996).
A recent FHWA guidebook identifies a range of demand modeling efforts and
summarizes the state of the practice in modeling methodologies (Cambridge
Systematics and BFA 1999). The success of these efforts has been limited, how-
ever, both by a lack of data and limitations in methodologies. To date, no com-
prehensive effort has been made to compare the preferences data developed by
various models and to assess their validity for widespread use in demand fore-
casting.
Existing Data Sources 29
FACILITIES
Data on pedestrian and bicycle facilities may describe the type of facility (side-
walk, shared-use path, on-road bike lane, pedestrian bridge, etc.), location,
length, width, physical condition, topography, intersection characteristics, and
other relevant features. Data on road facilities, such as number of lanes, lane
width, pavement quality, and intersection characteristics, can also be relevant to
analysis of bicycle and pedestrian travel.
Data on individual facilities need to be geographically referenced in some way
to be meaningful. This referencing may occur in a format as simple as a paper
map or a list of roads by jurisdiction. Increasingly, however, Geographic
Information Systems (GIS) are being used to maintain facility databases. GIS
can include databases in the form of lines (e.g., route segments) or points (e.g.,
intersections or bridges). GIS software packages provide a variety of analysis
and visual display capabilities that take advantage of the geographic nature of
the data.
Data on facilities can also be reported in summary formats. These might
include, for example, percentage of a city’s street network with continuous side-
walks, or miles of bike route by type and pavement condition within a city. An
example of aggregate reporting on road and highway facilities is the Federal
Highway Administration’s annual Highway Statistics (USDOT FHWA 2000).
Potential sources of data on bicycle and pedestrian facilities include:
● The U.S. Census Bureau’s Topologically Integrated Geographic Encoding
and Reference (TIGER) files,
● The National Transportation Atlas,
● The Rails-to-Trails Conservancy’s recreational trails database,
● State road databases, and
● Local road information.
The U.S. Census Bureau maintains its TIGER database, a digital database of
geographic features, including roads, covering the entire United States. The
database contains information about these features, such as location in latitude
and longitude, name, type of feature, address ranges for most streets, geo-
graphic relationship to other features, and other related information.
TIGER/Line files are publicly available and can be imported into most GIS soft-
ware packages.
30 Bicycle and Pedestrian Data
also a potential repository for other relevant information, such as the number,
characteristics, and trip patterns of trail users by trail as well as characteristics
of trail access and the surrounding area. If enough data of reasonable quality
could be assembled, this might provide the basis for analysis of factors influ-
encing both recreational and utilitarian nonmotorized travel.
Box 2-2
Statewide Bicycle Suitability Criteria
The Texas Transportation Institute (TTI) recently various bicycle suitability criteria is still in its
undertook a survey to determine the extent to inception. The majority of those states that had
which state Departments of Transportation bicycle suitability criteria in place had done so
(DOTs) have developed bicycle suitability cri- to meet state legislation that mandated their
teria for use in state roadway planning. The sur- formation and use as a part of a multimodal
vey revealed that 70 percent (11 of 16 sampled transportation plan. It appeared that the use of
states) had bicycle suitability criteria in place. traffic volume and lane width as primary suit-
The two most common criteria (one or both ability criteria was closely related to the fact
were used in every case) were the traffic vol- that this information was available in state
ume and the width of outside lanes or shoul- DOT databases. In addition to surveying cur-
ders. Thirty-five percent of the states with rent practice, the TTI report also makes recom-
suitability criteria also indicated that they mendations for developing and adopting
looked at heavy vehicles when considering bicycle suitability criteria.
traffic volume, 25 percent considered pave-
ment conditions, and 15 percent included traf- SOURCE: S.M. Turner, C.S. Schafer, and W.P. Stewart,
fic speed or speed limit criteria. Bicycle Suitability Criteria: Literature Review and State-
of-the-Practice Survey, Research Report 3988-1, prepared
by the Texas Transportation Institute, College Station, TX,
The conclusions from the survey indicate that,
1997, Internet: tti.tamu.edu.
with some exceptions, state implementation of
Increasingly, the word “crash” rather than “accident” is used to refer to inci-
dents that result in property damage, injury, or fatalities to road users. In the
case of pedestrians and bicycles, falls can also result in injuries and property
damage and should be included in safety nomenclature. Data on crashes and
falls can include the location of the incident; number and attributes of vehicles
and people involved; damage and injuries; characteristics of the incident loca-
tion; and contributing factors. Also related to safety, data on personal security
or crime is often relevant to bicycle and pedestrian travelers.
Crash and other safety data can be used:
● To identify trends by geographic area, facility type, severity, contributing
factors, etc.;
● To identify potential hazardous locations;
● To identify contributing factors to crashes and severity, including charac-
teristics of the individuals involved, vehicles, and environment;
● To identify potential countermeasures to reduce crashes;
● To evaluate the safety of various facility designs and operational policies;
● To identify crash costs; and
● To prioritize safety improvements.
Crash and other safety data are most useful when they can be related to expo-
sure data, a measure of the number of people at risk for a crash. Absolute num-
bers of crashes on a facility, for example, are not very meaningful unless they
can be related to the number of users of the facility. Dividing total crashes by
an exposure measure indicates the risk or likelihood of a crash per trip or dis-
tance traveled. A road with a high number of bicycle accidents and high bicy-
cle volumes may be safer (as measured by crashes per mile of travel) than a road
with a low number of bicycle accidents and low bicycle volumes.
The ability to associate crash data with facility data, such as shoulder width or
intersection characteristics, and with performance data for other modes, such
as traffic volumes and speeds, is also useful in crash investigation and safety
studies.
Sources of pedestrian and bicycle crash and safety data include:
● National crash and incident databases,
● National mortality and injury databases,
34 Bicycle and Pedestrian Data
NHTSA maintains two national crash databases for public use: the Fatality
Analysis Reporting System (FARS) and the National Automotive Sampling
System General Estimates System (NASS GES). The FARS is a database of all
fatal crashes involving motor vehicles on public roads. It contains over 100
attributes of the crash, vehicle, and people involved. GES data come from a
nationally representative sample of police-reported motor vehicle crashes of all
types, from minor to fatal. Approximately 90 data elements, collected from
police accident reports, are coded into a common format.
Both the FARS and GES databases are available from NHTSA, and the FARS
can be queried online. NHTSA also produces an annual summary report of
crash data entitled Traffic Safety Facts. This document includes summary sta-
tistics for pedestrian and bicycle crashes, such as fatalities and injuries by age,
gender, location (intersection or nonintersection), time of day, and related fac-
tors. NHTSA also produces a related series of traffic safety briefs, including lit-
erature on bicycling and pedestrians. BTS and the National Safety Council also
produce annual summary reports that include some data from the FARS and
GES (USDOT BTS 1998; NSC 1999).
Together, FARS and the GES have proven to be useful databases for tracking
trends and for national studies of crash characteristics, causes, and potential
countermeasures. Both databases can be used for research on pedestrian and
bicycle crashes that involve motor vehicles. Although FARS provides a com-
prehensive inventory of fatalities, deaths make up only a small proportion of
crashes. Data on nonfatal crashes are less comprehensive and reliable. The GES
suffers from potential sampling errors, and minor-injury and property-damage-
only crashes are typically underreported. Also, the GES data are limited by the
content and accuracy of the police reports from which they are obtained (see
section on police accident reports).
FARS and the GES also suffer from some specific limitations for pedestrian and
bicycle crash analysis. Neither database includes crashes that do not involve a
motor vehicle.3 Also, the databases do not include a number of variables that
may be important specifically in bicycle or pedestrian crash analysis. Examples
include whether the bicyclist was wearing a helmet, or whether objects were
present that might have obscured visibility. A lack of both appropriate crash
3According to a study by the Consumer Product Safety Commission, 10 percent of fatalities and 90
percent of injuries to bicyclists did not involve a motor vehicle (Rodgers et al. 1994).
Existing Data Sources 35
The National Center for Health Statistics (NCHS), run by the Centers for
Disease Control and Prevention, maintains databases that may be useful for
tracking pedestrian and bicycle injury trends. The National Vital Statistics
System (NVSS) includes annual reporting of all deaths in the United States, clas-
sified by cause (e.g., motor vehicle) and circumstance (collision with motor
vehicle, animal, bicycle, pedestrian, or fixed object, or fall). The National
Hospital Ambulatory Medical Care Survey (NHAMCS) includes a sample of
injuries in the United States classified by cause. Some summaries of NCHS data
on pedestrian and bicyclist injuries are published in the National Safety
Council’s annual Injury Facts (formerly Accident Facts).
One advantage of these sources over motor vehicle crash databases is that they
allow tracking of pedestrian and bicycle injuries and fatalities that do not
involve motor vehicles. Depending on the geographic resolution at which the
data can be analyzed, they could be used to develop local measures of risk (e.g.,
pedestrian injuries per capita). Also, the NHAMCS could potentially allow a
better analysis of injury characteristics and the resulting costs than is provided
by police accident reports. On the other hand, without linkages to specific crash
or location data the usefulness of these databases for crash investigation is lim-
ited or nonexistent. The authors are unaware of applications of these data to
pedestrian and bicycle injury analysis.
Another source of data on bicyclist injuries is the National Electronic Injury
Surveillance System (NEISS), maintained by the Consumer Product Safety
Commission. NEISS is based on a sample of hospitals that are statistically rep-
resentative of hospital emergency rooms nationwide. From the data collected,
estimates can be made of the numbers of injuries associated with consumer
products (including bicycles and bicycle helmets) and injuries treated in hospi-
tal emergency departments. Data are collected on a broad range of injury-
related issues, covering hundreds of product categories, and provide national
estimates of the number and severity of product-related injuries. The NEISS
data have been used in at least one study on bicycle-related injuries (Rodgers et
36 Bicycle and Pedestrian Data
Most states maintain, in varying forms, their own crash databases. NHTSA has
recently undertaken efforts to make state-level information more widely avail-
able and useful. One product of these efforts is the State Data System (SDS).
The SDS includes crash files from 17 participating states. The data are taken
from police accident reports and the specific data collected vary by state. Data
files are available from the individual states with state permission and a fee.
NHTSA has also produced a report from these data, State Data System: A
Summary of Motor Vehicle Traffic Crashes from State Crash Data Files, which
is available on its website. The summary report includes pedestrian fatalities
and injuries by age, location, and state.
Another NHTSA effort to make better use of existing state databases is the
Crash Outcome Data Evaluation System (CODES) project. CODES is a project
to link highway crash data to medical and financial outcome data. The linked
data were originally used to demonstrate the effectiveness of safety belts and
motorcycle helmets on death, disability, and costs. While linked data have not
been used specifically for bicycle and pedestrian crash analysis, linkage repre-
sents a potentially useful tool to conduct better crash analysis without collect-
ing additional data. By linking crash data to medical and financial records,
specific factors in bicycle and pedestrian crashes (helmet use, type of crash, loca-
tion, etc.) could be more closely linked to injury outcomes and the resulting
costs. This may assist with targeting resources for prevention and designing
appropriate countermeasures.
A similar effort is FHWA’s Highway Safety Information System (HSIS), which
currently contains state crash data for eight states—California, Illinois, Maine,
Michigan, Minnesota, North Carolina, Utah, and Washington. These crash
data are linked to roadway inventory files and traffic volume data. HSIS allows
users to analyze a large number of safety-related problems that can range from
basic “problem identification” issues to modeling efforts that attempt to predict
future accidents from roadway and traffic characteristics. HSIS is limited to
state roads, however, and does not cover local streets.
State-level databases can potentially provide a more complete sample of crash-
es than contained in the GES. As described above, they can be linked with data
in state road databases, such as facility type and traffic volumes. As a result,
they are generally more useful for local analysis and planning than national
crash databases. Not all state databases include crashes on local roads, howev-
er, and often do not provide enough detail for needed analyses. Also, as with
Existing Data Sources 37
the GES, the data are based on police accident reports and suffer from the same
limitations in reporting.
Police accident reports (PARs) serve as the basis for most local, state, and
national crash databases. Although many incidents are unreported, the majori-
ty of these involve only minor property damage and no significant personal
injury. PARs include specific categories of information on the crash location and
characteristics, as well as crash diagrams and narratives (compiled from witness
reports and follow-up investigation) describing what happened. The quality
and level of information collected varies locally.
PARs can provide a rich source of data for anyone with the resources to ana-
lyze them individually. However, they suffer from a number of limitations and
are frequently inadequate for remediation studies. These limitations include:
● Reporting formats and information vary from state to state (usually, each
state has a standard accident reporting form).
● Important variables are often not collected, and in particular many
variables relevant to pedestrian and bicycle crash analysis may not be
reported.
● Police officers may not be well trained in crash reporting and reports are
often filled out by civilians. As a result, the quality of reporting may vary.
In particular, injury severity is often estimated incorrectly.
● The reports are often available only in hard copy, although many depart-
ments are starting to adopt computerized reporting systems.
The widespread adoption of computer-based reporting systems has the poten-
tial to overcome many of these limitations and provide better and more useable
crash data. These opportunities are discussed in chapter 4.
In addition to crashes and falls, crime can represent a safety hazard for pedes-
trians and bicyclists. Actual and perceived risks to personal security may in
some cases be a strong deterrent to pedestrian or bicycle travel. Data on the
locations and characteristics of criminal incidents, such as muggings, can help
local authorities improve enforcement in needed areas and reduce risks to trav-
elers. While some aspects of crime prevention are beyond the purview of trans-
4 This section is based on personal communication with Richard Blomberg, Dunlap & Associates,
Stamford, CT.
38 Bicycle and Pedestrian Data
portation planning agencies, in many cases, site and facility design can influence
the safety of bicyclists and pedestrians. Examples of important design features
include lighting, placement of vegetation, sightlines, and orientation of build-
ings.
Data on incidents of crime are primarily maintained at the city level in police
crime logs. Many of the same statements that apply to local crash reporting
apply to the reporting and management of crime data. For example, cities are
developing—to varying degrees—GIS-based systems that can display the loca-
tions of incidents as well as store the various attributes of these incidents. In this
outreach effort, no particularly outstanding issues regarding security-related
data for bicycle and pedestrian analysis were identified, but nonetheless this is
an area that should not be overlooked. Perceptions of safety are also important,
but may not correlate to the reported level of crime.
The Transportation Equity Act for the 21st Century (TEA-21) calls for “a
national accounting of expenditures and capital stocks (facilities and vehicles)
on each mode of transportation and intermodal combination.” Ideally, this
would include estimates of the current value of sidewalks, shared-use paths,
pedestrian bridges, etc., but the data currently do not exist.
Tracking of bicycle stocks is one way of tracking trends in bicycle usage. The
following sources on bicycle expenditures and stocks are known:
● The Bicycle Manufacturer’s Association has data on bicycles sold from
1970 to the present in five-year increments. These data are available in the
BTS report, National Transportation Statistics, which is produced
annually (USDOT BTS 1999).
● The Bureau of Labor Statistics’ (BLS’s) Consumer Expenditure Survey
includes data on personal expenditures on bicycles. These expenditures
can be tabulated by age, education, income, household composition,
occupation, region of residence, and other variables. Microdata CD-
ROMs that include bicycle expenditure information can be purchased
from BLS.
● Purchasing data are available from industry surveys, such as a 1990 sur-
vey by Rodale Press of 3,200 new bicycle purchasers. These data are not
currently reported on a regular basis.
● The Federal Transit Administration’s (FTA) grant management system is
a potential source for FTA funding of bicycle-transit links. However, for
5 This section is partially based on unpublished information from Paul Schimek, Cambridge, MA.
Existing Data Sources 39
most grants that include funds for bicycle improvements, such as for the
purchase of racks and lockers, the system does not separately identify this
information. Likewise, FHWA tracks bicycle and pedestrian funding for
stand-alone projects—that is, where the funding is not part of a larger
transportation project. The information covers projects that use federal
transportation money (i.e., ISTEA, TEA-21), and does not show spending
categories.
SECONDARY DATA
Various research studies have been conducted on factors such as the safety
effects of design features and the demand impacts of various design features and
other policies. Results are frequently disseminated through publication of
reports and conference proceedings, and may be used to inform the develop-
ment of design manuals. Local planners and advocates, however, often do not
have the time or the resources available to locate published studies. Also, many
design features—particularly recent innovations, or those that have seen little
implementation—have not been thoroughly studied. A further complication, in
the case of demand impacts, is that the relative impacts of various policies and
actions may work together in ways that are difficult to untangle. As a result,
applying the results of demand studies to local situations can be very difficult.
The U.S. Department of Transportation’s National Bicycling and Walking Study
produced a series of reports in the 1990s, some of which included references to
research results on various topics. A forthcoming publication by NHTSA will
40 Bicycle and Pedestrian Data
Manuals of Practice
REFERENCES
Ashley, C.A. and C. Banister. 1989. Cycling to Work from Wards in a Metropolitan
Area. Traffic Engineering and Control 30:6–8, June–September.
Existing Data Sources 41
3
SUMMARY OF DATA NEEDS AND PRIORITIES
43
44 Bicycle and Pedestrian Data
Perhaps of highest priority, there is a strong need for comprehensive and sys-
tematic data on usage, including potential usage. Even basic information, such
as total travelers, trips, and distance traveled by bicycle or foot, does not exist
in most areas. Nearly 40 percent of respondents to the outreach effort identi-
fied a need for better data of this type. In particular, respondents expressed three
related but distinct needs:
● Overall indicators of usage and trends in usage, such as numbers of peo-
ple who bicycle or walk; total numbers of bicycle or walk trips; mode
shares; and miles of travel by nonmotorized modes. The need for such
indicators was expressed primarily at the local level but also at the state
and national levels.
● Counts (traffic volumes) on specific facilities, in some cases by character-
istics such as time of day, day of week, or type (e.g., work, shopping,
recreational).
Data Needs 45
Table 3-1
Assessment of Data Priorities
Priority
Quality of for better
Data type Description Primary uses existing data data
Usage, trip, Number of Conditions and trend analysis Poor High
and user bicyclists and
characteristics pedestrians by Network planning (prioritize
facility or improvements)
geographic area
Crash analysis (develop exposure
measures)
Data on user preferences, attitudes, and expressed needs of existing and poten-
tial bicyclists and pedestrians appear to be a secondary but still important pri-
ority.1 Fourteen percent of respondents expressed a desire for better data in this
area. Respondents specifically requested data on relative preferences for facility
types, to assist with both facility design and travel modeling; reasons why peo-
ple do not bicycle or walk and what would encourage them to do so; and per-
ceptions of safety and the influence of safety concerns on bicycling or walking.
A fair amount of information on expressed preferences—as gathered from var-
ious local and national surveys—already exists, and has to some extent been
summarized through the National Bicycling and Walking Study. These data
provide a general sense of what factors are important in a person’s decision to
walk or bicycle. Recent research to establish a Bicycle Compatibility Index is
also helping establish relative preferences for specific design features (USDOT
FHWA 1998). Knowledge of user preferences falls short, however, when it
comes to modeling the impacts of specific policies and facility design character-
istics on travel choices. This is because there are many factors that influence
these decisions, and they can interact in complicated ways. Advances in demand
modeling as well as better data to support this modeling are therefore required
to further quantify relative preferences.
Fortunately, a systematic and comprehensive data collection effort on user pref-
erences, similar to a system of tracking usage and crash data, is probably not
necessary. There is strong evidence from travel modeling experience that pref-
erences are to a large extent “transferable” from one area to the next, and that
it is therefore not necessary to perform preference surveys in every city or met-
ropolitan area.
Current challenges are:
● To further disseminate results (including recent surveys and studies) on a
national basis,
● To compare results across geographic areas to determine their transfer-
ability,
● To synthesize results in a manner that is most useful to local practition-
ers, and
● To identify additional research efforts that are required to support specific
purposes such as demand modeling and facility design.
1Note that it is important to know preferences for potential as well as actual bicyclists and pedestri-
ans, therefore the term “user” is applied loosely to indicate all people for whom bicycle and
pedestrian-related policies could potentially impact travel choices.
48 Bicycle and Pedestrian Data
FACILITIES
Better data on the location, characteristics, and extent of bicycle and pedestri-
an facilities was identified as a need by 15 percent of respondents to the out-
reach effort. As with preferences data, improved data of this type would be
useful but are probably not as high a priority as basic usage data. For local
planning purposes, even if a comprehensive inventory of facilities does not exist
(and therefore total mileage by type, condition, etc., cannot be quantified), any-
one familiar with an area has a general sense of what conditions are like and
where specific problems might exist. It is easier to assess existing facility condi-
tions through observation than to obtain “hard” data such as user counts and
surveys. Local jurisdictions, for example, frequently conduct inventories of
bicycle facilities for the purpose of developing bicycle plans.
Nevertheless, both needs and opportunities exist for improving facilities data.
Better facilities data would appear most useful for local planning efforts. A
number of respondents identified a desire for databases indicating locations and
characteristics of specific facilities. Potential uses include developing compati-
bility ratings and recommended routes, identifying deficiencies, and identifying
and prioritizing improvements. Local data on facility characteristics would also
assist in crash studies and in forecasting travel demand.
A few respondents also identified a desire for aggregate-level reporting of facil-
ities data (e.g., miles of bicycle paths per capita in a city, county, or metropoli-
tan area). Tracking the extent of facilities at the metropolitan, state, and
national levels would be useful in helping planners and policymakers assess the
extent to which adequate facilities are provided.
Particular needs for improving facilities data include:
● Incorporating attributes relevant to bicycle and pedestrian travel (e.g.,
sidewalk and crosswalk locations) into roadway databases;
● Expanding roadway databases to include bicycle- and pedestrian-specific
links, such as off-road trails;
● Incorporating other environmental and land use variables that relate to
the quality of an area for bicycling and/or walking; and
Data Needs 49
SECONDARY DATA
Better secondary data appear to be a relatively high priority for planners, engi-
neers, consultants, and advocates at the local level. Important types of second-
ary data include research results on the safety, demand, and other impacts of
design features and policies; recommended design practices; and model policies
and programs relating to design, education, and promotion.
The safety and user-preference impacts of various design features, especially for
bicycle travel, were identified by about 30 percent of outreach respondents and
appear to be the highest priority need. While recent research has focused on a
number of bicycle and pedestrian design issues, many issues have not been thor-
oughly researched, and evidence on the appropriateness of various treatments
is often conflicting or nonexistent. Furthermore, existing research results may
50 Bicycle and Pedestrian Data
REFERENCES
Cambridge Systematics and Bicycle Federation of America (BFA). 1999. Guidebook
on Methods to Estimate Non-Motorized Travel, FHWA-RD-98-165. Produced
for the U.S. Department of Transportation, Federal Highway Administration.
Washington, DC.
National Academy of Sciences (NAS), Transportation Research Board (TRB). 1997.
Information Needs to Support State and Local Transportation Decisionmaking
into the 21st Century, Proceedings of a Conference, Irvine, CA, March 25.
Washington, DC: National Academy Press. Internet: www.bts.gov/needs as of
May 5, 2000.
U.S. Department of Transportation (USDOT), Bureau of Transportation Statistics
(BTS). 1998. Transportation Statistics Beyond ISTEA: Critical Gaps and
Strategic Responses, BTS98-A-01. Washington, DC.
U.S. Department of Transportation (USDOT), Federal Highway Administration
(FHWA). 1998. Development of the Bicycle Compatibility Index: A Level of
Service Concept, Final report, FHWA-RD-98-072. Prepared by Turner-Fairbank
Highway Research Center.
Options for
CHAPTER
Addressing Data
Needs 4
GENERAL RECOMMENDATIONS
51
52 Bicycle and Pedestrian Data
1 This need was identified in the development of the Guidebook on Methods to Forecast Bicycle and
Pedestrian Travel (Cambridge Systematics and BFA 1999).
Options for Addressing Data Needs 55
(GPS) units to track trips should also be monitored and potential appli-
cations to monitoring pedestrian and bicycle travel explored.
● Improve the state of the practice in bicycle and pedestrian demand fore-
casting. Data on forecast usage can be at least as important as data on
existing usage for many planning and design applications and for sup-
porting funding allocation decisions. Forecasts may be important in
designing a trail to ensure adequate width, for example, or in planning a
network of routes to maximize usage.
permit the transfer of crash data into central databases, eliminating the
need for manual data entry. As a result, data on all reported crashes can
be made readily available for analysis.
● Geographic information systems are facilitating the management, display,
and linkage of data. GIS will assist researchers and planners in linking
data to facility and environmental databases, and will assist lay people
through its capabilities of visually displaying information. GIS-based sys-
tems for maintaining crash records, facility information, and other data
are in various levels of implementation by states, metropolitan agencies,
and local jurisdictions.
● Global positioning systems are permitting the recording of precise crash
locations.
● National Highway Traffic Safety Administration (NHTSA) is undertaking
efforts to link crash and medical records databases in some states. Use of
these data for bicycle and pedestrian crash analysis, however, will require
funding of specific research studies.
● NHTSA is also developing crash-typing software. The Pedestrian and
Bicycle Crash Analysis Tool (PBCAT) helps users create a database of
their pedestrian and bicycle crashes from hardcopy accident reports, helps
users capture detail on roadway and environmental conditions at the time
of the crash, types/categorizes the crashes, helps users produce some basic
tables and graphs, and links the crash types to some countermeasures.
In implementing new technologies for crash reporting, analysis, and data
management, consideration should be given to specific needs for bicycle and
pedestrian crash reporting. The potential of computer-based crash reporting
systems, for example, should be leveraged for bicycle and pedestrian crash
analysis through the development of expert systems to guide data entry.
Expert systems can request specific information keyed to the nature of the
crash. For example, if a bicyclist were involved, the computer would prompt
to determine whether or not the bicyclist was wearing a helmet.
Some specific recommendations to improve pedestrian and bicycle crash
data include:
● Build a consensus on characteristics that are important to record in
reporting crashes involving bicyclists and pedestrians; incorporate these
characteristics in standards for computer-based reporting systems. Crash
reporting systems should be designed with input from bicycle and pedes-
trian safety researchers to include important information relevant to bicy-
cle and pedestrian crash analysis. To maximize its usefulness, a standard
Options for Addressing Data Needs 59
SECONDARY DATA
Needs for secondary data may be addressed through more widespread dis-
semination of existing research results; through additional research on
design features, policies, etc. that have not been fully studied; and through
the development and dissemination of recommended practices as well as case
studies of successful programs and applications. Some specific recommenda-
tions include:
● Continue to improve the availability of existing research. The National
Transportation Library, the National Bicycling and Walking Study, and
the Bicycle and Pedestrian Clearinghouse have all made strides in this
direction and the Internet provides excellent new opportunities in this
2 Richard Blomberg of the Transportation Research Board Pedestrian Committee has been investigat-
ing opportunities and constraints to the adoption of new reporting systems. Involvement by the
NHTSA may be helpful in promoting widespread agreement on and adoption of such systems.
3 Self-reporting through surveys has previously been used to estimate the frequency and types of non-
motor-vehicle crashes and incidents (Moritz 1997).
60 Bicycle and Pedestrian Data
REFERENCES
Cambridge Systematics and Bicycle Federation of America (BFA). 1999. Guidebook
on Methods to Estimate Non-Motorized Travel, FHWA-RD-98-165. Produced
for the U.S. Department of Transportation, Federal Highway Administration.
Washington, DC.
Miller, J.S. 1997. Using Technology to Help Overcome Institutional Obstacles to
Improved Crash Records Processing. Transportation Research Record 1581.
Moritz, W.E. 1997. A Survey of North American Bicycle Commuters: Design and
Aggregate Results, presented at the Transportation Research Board Annual
Meeting, Paper No. 970979. Washington, DC.
Results of Outreach
APPENDIX
Effort
A
METHODOLOGY
61
62 Bicycle and Pedestrian Data
Table A-1
Email Listservs Contacted
In addition, to obtain broad coverage (in particular of local planners and advo-
cates), the questionnaire was sent to 10 email listservs (see table A-1).
The questionnaire was not intended to yield a random or representative sample
of responses. Rather, it was viewed as a low-cost means of giving as many peo-
ple as possible the opportunity to respond and provide feedback. In conjunc-
tion with the other sources used for this study, it helps provide a picture of the
most pressing data needs for various types of activities.
RESPONSES
DATA NEEDS
Table A-2
Survey Responses by Type of Respondent The outreach effort conducted for this
project asked the following questions to
Type of respondent Responses ascertain data needs:
Project/system planning
State (bike/pedestrian coordinator) 8
● What types of data would be most
Metro (metropolitan planning useful to you but are not available?
organization) 8
● How would these data be useful?
Local government 8
Consulting 5 Responses were categorized according to
Subtotal 29
Advocacy
the type of data mentioned. Response fre-
National 3 quencies are shown in table A-5. Since the
State/local 17 responses do not represent a random
Subtotal 20 sample of bicycle and pedestrian plan-
Research ners, researchers, and advocates, the fre-
Safety/design 6 quency of responses should not be viewed
System planning/demand analysis 3
Health/physical activity 1
as a prioritization. In particular, most
Subtotal 10 responses came from planners and advo-
Health/physical activity cates at the state and local level. The
State department of health 7 results are generally consistent, however,
Other 3 with data needs identified through other
Subtotal 10 forums by researchers and by national-
Other/unknown 9
Total 78
level planners, policymakers, and advo-
cates. Results are discussed according to
the major categories of identified data
needs as shown in table A-5.
Nearly 40 percent of survey respondents identified a need for better data on sys-
tem usage and user characteristics. As discussed above, the most widely
accessed sources of these data are the decennial census, the NPTS, and metro-
politan household travel surveys. Other counts and surveys are conducted spo-
radically in particular areas or nationwide. The most frequently mentioned
needs include:
● Demand/usage indicators. Sixteen respondents expressed the need for bet-
ter indicators of system usage. Such indicators might include total bicy-
cling or walking trips, numbers of people who bicycle or walk, mode
shares, and miles of travel by nonmotorized modes. The need for such
indicators was expressed primarily at a local level (city, county, or metro-
politan area) but also at the state and national geographic scales. Reasons
for wanting these data include tracking trends, determining the effective-
Results of Outreach Effort 65
Table A-5
Outreach Effort: Summary of Identified Data Needs
Secondary data
Facility design 22 28
Safety impacts 13 16
Demand impacts 7 9
Recommended practices 4 5
Costing data for planning 2 3
Effects on motorist behavior and attitudes 1 1
Policies and programs 10 13
Recommended/model practices 6 8
Benefits of various policies/programs 3 4
Safety impacts 3 4
Demand impacts 1 1
Results of Outreach Effort 67
indicators,” but with the difference that the data are desired for specific
facilities rather than for an area as a whole. Counts, if performed in a sys-
tematic manner at an adequate sample of locations, can provide a basis
for estimating aggregate usage and tracking trends in usage; developing
exposure measures for crash analysis; and calibrating travel models. They
can also directly assist planning activities such as designing adequate facil-
ities to accommodate traffic and prioritizing facilities for improvement.
● Exposure data. Seven respondents specifically indicated a need for “expo-
sure” data. Exposure data are similar to usage indicators in that they
quantify trips, miles of travel, etc. An additional purpose of exposure
data, however, is to develop crash rates or risk measures (e.g., crashes per
bicycle-mile of travel). As a result, data used to develop exposure meas-
ures may have more specific requirements than data that are collected just
to determine levels and trends in system usage. It may be desirable, for
example, to collect exposure data by facility type, time of day, bicyclist
experience level, or any other measure that is important in classifying and
analyzing crashes.
● Trip generation data. Five respondents (all consultants) identified the
need for bicycle trip generation data by type of land use. Automobile trip
generation rates, expressed in terms of trips per thousand square feet or
another measure of the size of development, are widely used in trans-
portation planning. They are developed by regressing traffic counts
against size of development and other related measures, based on surveys
at various locations, and are reported in the Institute of Transportation
Engineers’ Trip Generation. Consultants who work with such data would
clearly appreciate similar data for bicyclists.
● Other usage-related data. Other desired data related to system usage
include bicycles in use or owned (three respondents) and helmet usage
(three respondents).
Facilities
Twelve respondents identified a need for better data on bicycle and pedestrian
facilities. Two-thirds of these identified a need for databases indicating locations
and characteristics of specific facilities, that is, “disaggregate” facility data.
Desired data include not only the presence and characteristics of bicycle and
pedestrian facilities, but other relevant characteristics such as roadway lane
widths and traffic volumes. A number of respondents noted that having these
data in Geographic Information System (GIS) format would be of great help in
planning and analysis.
Better facilities data would primarily be useful in local planning activities,
including developing compatibility ratings and recommended routes, identify-
ing deficiencies, and identifying and prioritizing improvements. Some respon-
dents also expressed a desire for facilities data in a format that can be related to
other datasets, such as travel volumes and crashes.
In addition to facility databases, four respondents noted that it would be useful
to have summary statistics of miles of bicycle lanes, sidewalks, etc. The purpose
of having these data would be to measure progress in providing facilities and to
set “benchmarks,” for example, an actual versus desired ratio of bike lane-miles
per capita or per road-mile.
Three respondents noted that it would also be useful to have either general
inventories or know the specific locations of ancillary facilities, including bicy-
cle parking and shower/locker facilities at workplaces.
Crashes
quality data while others might not. Some jurisdictions, for example, have
developed GIS-based crash databases, but this technology has not yet
been adopted throughout the country. The ease with which crash data can
be obtained from the responsible agency can also vary locally.
● Crash characteristics and contributing factors. A number of respondents
also indicated a need for additional and better quality data collection on
crash characteristics and contributing factors. These data would help
inform crash analysis and countermeasure development. Specific areas in
which better data are needed include environmental variables such as
roadway geometrics and visibility impairments; characteristics of those
involved such as demographics, helmet use, light use, and bicyclist skill
level; and other potential causal or contributing factors.
● Specific types of crashes. Some respondents also noted specific types of
crashes for which data are particularly lacking. While the Fatal Accident
Reporting System (FARS) provides a comprehensive inventory of fatali-
ties, injury databases may not be comprehensive, and as the severity of the
crash decreases the quality of the data and likelihood of reporting also
decreases. Furthermore, a general lack of data on crashes or incidents that
do not involve motor vehicles was noted. These might include, for exam-
ple, bicycle-pedestrian collisions or single-bicycle incidents. Data of this
type, in addition to tracking trends, would be useful in analyzing the safe-
ty impacts of various facility designs or specific facilities.
● Injuries and outcomes. Two respondents noted the need for better data on
injuries and medical outcomes of crashes. Currently, injury severity is
coded by an often untrained crash witness. Better data on injury types,
severity, and costs can assist in analyzing the full costs of various types of
crashes, designing appropriate countermeasures, and prioritizing specific
sites or crash types for countermeasures.
● Summary data. Seven respondents indicated a desire for better summary-
level data on crashes. Specific requests were for better reporting of con-
tributing factors and causes of bicycle and pedestrian crashes; more
complete reporting of nonfatality and nonmotor-vehicle crashes; bicyclist
violations of the motor vehicle code; and crime on trails.
Secondary Data
The data needs discussed above all involve the collection, dissemination, and/or
summary analysis of primary data. Over one-third of survey respondents, how-
ever, also interpreted “data” to mean secondary data sources, such as results of
research studies and recommended practices. Respondents involved in local
planning and facility design rely heavily on these data to assist in choosing
70 Bicycle and Pedestrian Data
In addition to asking about existing sources and data needs, the outreach effort
conducted for this project also asked respondents whether they had any sug-
gestions for:
● More effective dissemination and utilization of existing data sources,
● Cost-effective ways of collecting new data (enhancement of existing sur-
veys, use of advanced surveillance technologies, etc.), and
● Assistance activities that would help encourage and facilitate the collec-
tion of local data.
Responses are summarized below.
Results of Outreach Effort 71
Half of the survey respondents provided suggestions for making better use of
existing data. Not surprisingly for a survey conducted primarily by email, the
vast majority of these respondents commented that all available data should be
made readily accessible via the Internet. (While use of the Internet is not yet
ubiquitous, one respondent commented that “Almost all of us who do bike
planning for a living are online and in contact with each other.”) A few respon-
dents specifically suggested that a centralized, searchable database would be
useful, as would links to related sites with relevant data such as NHTSA and
the Centers for Disease Control and Prevention (CDC). A few respondents also
noted existing good examples of data dissemination, including the Bureau of
Transportation Statistics’ (BTS’) National Transportation Library and the
Federal Highway Administration (FHWA) Bicycle/Pedestrian Clearinghouse,
and suggested expanding these applications.
Five respondents noted a general need for better awareness of existing data,
where to obtain it, and how to use it.
Other suggestions for collecting and disseminating existing data included:
1. Use of existing publications, such as American Planning Association
newsletters and the Institute of Transportation Engineer’s Trip
Generation;
2. Collection and dissemination of local data and survey results through
state bicycle and pedestrian coordinators or other state-level staff;
3. Dissemination through other groups, such as the National Bicycle Safety
Network and state departments of health;
4. Additional analysis and publication of summary-level bicycle- and pedes-
trian-related crash data, using existing national and state databases;
5. More widespread publication and presentation of bicycle- and pedestrian-
related research;
6. Maintaining contact lists of local governments and advocacy groups who
are working to improve bicycle and pedestrian conditions; and
7. Sponsorship of a conference on bicycle and pedestrian site design issues,
including case studies, in conjunction with developers.
nologies and 2) improved surveying and sampling methods. Suggestions for use
of new technologies included:
● Video monitoring for bicycle and pedestrian counts—including both new
monitoring systems and use of existing systems established for traffic or
street security monitoring;
● Infrared sensors;
● Dataloggers that are placed on a person and measure physical and envi-
ronmental variables;
● Use of GPS for pedestrian trip tracking; and
● Laptop-based crash reporting systems.
Suggestions for conducting or expanding surveys included:
● Addition of questions to the NPTS;
● Addition of questions to health surveys, including the Behavioral Risk
Factor Surveillance System and the Youth Risk Factor Survey;
● Distribution of surveys at bicycle shops and pedestrian trip generators,
such as schools and shopping malls;
● Surveys of people accessing mass transit;
● Surveys using the Internet or newspapers; and
● Use of local bicycle groups to identify problem areas.
Respondents’ suggestions for how the BTS or other federal agencies could assist
with data collection generally fell into three categories: technical assistance,
financial assistance, and regulations/ mandates.
● Technical assistance. Seven respondents identified the need for technical
assistance, specifically:
- Developing uniform data collection methodology, reporting for-
mats, and analysis methods;
- Demonstrating, evaluating, and testing new data collection tech-
nologies; and
- Developing case studies of successsful data collection efforts.
● Financial assistance. Eleven respondents noted a shortage of local
resources for data collection and suggested that financial assistance would
Results of Outreach Effort 73
Assessments of Data
Needs B
T his project also reviewed the results of other recent assessments of
bicycle and pedestrian data needs. These include:
● A 1997 conference sponsored by the Bureau of Transportation
Statistics (BTS) in Irvine, California, to identify transportation-
related information needs, including bicycle and pedestrian data;
● The BTS’ Transportation Statistics Beyond ISTEA: Critical Gaps and
Strategic Responses, an assessment of needs for transportation statis-
tics; and
● A Federal Highway Administration (FHWA)-sponsored guidebook
on methods for forecasting bicycle and pedestrian travel.
To a large extent, the recommendations from these efforts corroborate
the results of the outreach effort conducted for this project. Key findings
are summarized below.
IRVINE CONFERENCE
75
76 Bicycle and Pedestrian Data
REFERENCES
Cambridge Systematics and Bicycle Federation of America (BFA). 1999. Guidebook
on Methods to Estimate Non-Motorized Travel, FHWA-RD-98-165. Produced
for the U.S. Department of Transportation, Federal Highway Administration,
Washington, DC.
U.S. Department of Transportation (USDOT), Bureau of Transportation Statistics
(BTS). 1998. Transportation Statistics Beyond ISTEA: Critical Gaps and
Strategic Responses, BTS98-A-01. Washington, DC.